Kirkendall Porosity Research Articles

Microstructural evolution of brazing 422 stainless steel using the BNi-3 braze alloy

The 422 stainless steel (422SS) is one of the typical martensitic stainless steels with both excellent creep strength and corrosion resistance up to 650°C. Its application includes steam turbine blades, high temperatures bolting ... etc. Repair welding of 422SS is one of the most common methods to fix the turbine blade. However, repair brazing of surface shallow cracks, e.g., less than 1 mm in depth, is an alternative way to fix such blades. The microstructural evolution of brazing 422SS with BNi-3 braze alloy using both infrared and furnace brazing was performed in the study. Based on the experimental results, BNi-3 cannot effectively wet 422SS substrate below 1025°C. As the brazing temperature increases above 1050°C, comprehensive wetting can be obtained in 1200 sec. For the infrared brazed specimen with a short brazing time, the cooling path starts from the formation of a BNi3 phase in the molten braze, subsequently forms a Ni-rich phase, and finally a eutectic phase is solidified from the residual eutectic liquid. The microstructure of the furnace-brazed specimen is similar to that of infrared brazed specimen, but the interfacial reaction zone is significantly increased in furnace brazing. There are Kirkendall voids in the braze close to the interface between BNi-3 and 422SS, and the size of Kirkendall porosity is increased with increment of the brazing time and/or temperature. The homogenization treatment of the brazed joint at 900°C results in growth of both the interfacial reaction zone and porosity.

Microstructure of Al 3Sc with ternary transition-metal additions

The microstructure of binary Al 3Sc and ternary Al 3(Sc 1− y X y ), where X is one of the transition metals from Group IIIA (Y), IVA (Ti, Zr or Hf) or VA (V, Nb or Ta), was investigated as a function of alloying element concentration for 0.1≤ y≤0.75. Alloys with Group IIIA and IVA additions exhibited a single L1 2 solid-solution phase with some Kirkendall porosity. At the highest concentration studied, a second phase precipitated with the D0 19 (Y), D0 22 (Ti) or D0 23 (Zr and Hf) structure. Conversely, alloys with Group VA additions exhibited both the L1 2 trialuminide phase and a dendritic trialuminide second phase with D0 22 structure for all concentrations studied. The solubility limit in the ternary L1 2-type Al 3(Sc 1− y X y ) phase was high for Group IIIA and IVA metals (almost 12.5 at.% or y=0.5), and much lower for Group VA metals (from about 1.8 at.% or y=0.07 for Ta to about 2.7 at.% or y=0.11 for V). Similarly, the solubility limit of Sc in the non-L1 2 phases decreases from the Group IIIA trialuminide to the Group VA trialuminides. The lattice parameter of the L1 2 solid-solution decreased linearly with increasing concentration of Group IVA and VA metals, but increased linearly with concentration of Y (Group IIIA). This linear concentration dependence of the lattice parameter is found to correlate with the atomic size mismatch between Sc and the transition metal. The microhardness of the L1 2 solid-solution increased linearly with increasing concentration of ternary elements. The concentration dependence of hardness is strongest for Group VA metals and weakest for Group IVA metals, for which a correlation is found with the concentration dependence of lattice parameter.

Passivation of Nickel-Base Superalloy Inconel 690 by Pack-Cementation Chromium Coatings

French Pressurised Water Nuclear Reactors have steam generators made of Inconel 690, a nickel-base superalloy with high chromium and iron contents. Corrosion of this superalloy in primary water conditions may be reduced by a surface treatment of the tubes. The deposition of chromium onto Inconel 690 in cementation packs containing a chromizing mixture and a CrCl 3 activator is studied between 1273 K and 1523 K in sealed quartz tubes. At constant volume, liquid CrCl 2 is synthesised in the conditions used. Two layers are observed at 1273 K and 1323 K by metallographic observations. The composition of the outer layer is a thick a Cr-rich layer with 4.6 at% Ni and 1.7 at.% Fe. An inner diffusion layer also grows. It corresponds to a γ-Cr-Ni layer. At 1523 K, only the Cr-rich layer still grows and Kirkendall porosity appears. Growth rates of the Cr-rich layer are parabolic whatever the temperature. The rate-limiting step is very likely diffusion of chromium into the chromium-rich layer. Finally, oxidation tests were carried out in a loop simulating primary water conditions of Pressurized Water Reactor.

On the diffusion of aluminium and titanium in the Ni-rich Ni–Al–Ti system between 900 and 1200°C

Interdiffusion coefficients in the Ni-rich portion of the Ni–Al–Ti system have been determined, in the temperature range 900–1200°C, concentrating particularly on the face-centered cubic (γ) phase. The values were derived from concentration profiles measured by electron probe microanalysis, using a modified form of the Boltzmann–Matano method. The diffusion couples were of three types: γ/γ, γ/γ′ and γ′/γ′, where γ′ denotes the L1 2 phase. The two major coefficients D ̃ Ni,γ AlAl and D ̃ Ni,γ TiTi are very nearly equal; moreover these vary only slightly with composition. The minor coefficients D ̃ Ni,γ AlTi and D ̃ Ni,γ TiAl are up to an order of magnitude smaller and show considerable composition dependence. When plotted on the ternary phase diagram, the D ̃ Ni,γ AlAl and D ̃ Ni,γ AlTi contours lie approximately parallel to the Al iso-concentration lines whereas the D ̃ Ni,γ TiTi and D ̃ Ni,γ TiAl contours follow the iso-concentration lines for Ti. The activation energies of all coefficients were found to be similar at the reported compositions. Observations are made concerning the accumulation of Kirkendall porosity in the interdiffusion zone, the locations of the γ/(γ+γ′) and γ′/(γ+γ′) phase boundaries and also their rate of migration. Data for the interdiffusion of Al and Ti on the Al/Ti site of the intermetallic phase γ′ are also reported, and the values are rationalised in the context of the diffusion mechanism which is known to operate in that phase. The kinetic data enable reasonable predictions to be made of the parabolic rate constant pertaining to the γ/γ′ interfaces, implying that local equilibrium is indeed adopted during migration under diffusional control.

Sintering of MnZn-ferrite powders prepared by hydrothermal reactions between oxides

Abstract The sintering of ferrite powders prepared by a hydrothermal treatment of starting oxides was studied in comparison with conventional pre-sintered powders. Two hydrothermal powders were prepared, one with a similar degree of conversion into spinel products as the pre-sintered powder but with a larger specific surface, and the other with a lower specific surface area and a high degree of conversion into MnZnFe2O4. In the first powder, non-homogeneity caused primarily by the agglomeration of the Mn-rich particles (Mn3O4 and ZnMn2O4) resulted in the formation of Kirkendall porosity, which significantly influenced the densification process.

Interdiffusion and Kirkendall Effect in Doped BaTiO3–BaZrO3 Perovskites: Effect of Vacancy Supersaturation

Interdiffusion in Sc‐doped and Ta‐doped BaTiO3‐BaZrO3 perovskites was examined, both theoretically and experimentally. Analytical expressions for the lattice velocity, v, and the interdiffusion coefficient,D̃, were obtained. The v and the D̃ were numerically evaluated as a function of dopant type (trivalent‐acceptor or pentavalent‐donor) and concentration using literature values of oxygen vacancy diffusivity, DVO, the A‐site vacancy diffusivity, DVA, and assumed values for the B‐site vacancy diffusivities, DVB and DVB′. Regardless of the chosen values of DVB and DVB′, the calculated D̃ increased with the Ta concentration and decreased with the Sc concentration. The dependence (shape) of the calculated v vs dopant type and concentration, however, was a function of the magnitudes of DVA and DVB. For DVA < (DVB, DVB′) < DVO, the calculated v exhibited a maximum at ∼0.8% acceptor dopant. Diffusion couples in the BaTiO3‐BaZrO3 system doped with either Sc (acceptor) or Ta (donor) were annealed in air over a temperature range between 1300° and 1500°C. Extensive Kirkendall porosity was observed in several samples. The measured porosity vs dopant concentration (acceptor/donor) trace was similar to that of calculated v vs dopant concentration (acceptor/donor). However, the measured D̃ vs dopant concentration (acceptor/donor) was not in accord with the calculations. This difference is attributed to a nonequilibrium vacancy supersaturation and the accompanying porosity formation. To obtain further evidence of the deviation from equilibrium, experiments were also conducted on both coarse‐grained and fine‐grained samples. The coarse‐grained samples, which are expected to have fewer vacancy sinks and sources, exhibited greater Kirkendall porosity and greater D̃ values compared to the fine‐grained samples, consistent with a greater deviation from equilibrium.

Diffusion bonded Mokumé Gane decorative metal laminates

Vacuum diffusion bonding has been used successfully to fabricate a range of new metal combinations suitable for Mokume Gane use. New metal systems of mutual solubility, mutual insolubility, and highly reactive couples can be used, greatly extending the range of colours available. Kirkendall porosity occurs during the fabrication of many material couples, but this is normally closed during subsequent rolling. If brittle intermetallics form at the laminate interfaces, delamination may occur during forging and rolling. However, with care, usable laminates can be rolled to sheet. Laminates which contain metals of very different mechanical properties undergo internal fracture during rolling, and this is shown to be the result of internal tensile stresses generated during the rolling operation.

Diffusion bonded Mokumé Gane decorative metal laminates

Vacuum diffusion bonding has been used successfully to fabricate a range of new metal combinations suitable for Mokume Gane use. New metal systems of mutual solubility, mutual insolubility, and highly reactive couples can be used, greatly extending the range of colours available. Kirkendall porosity occurs during the fabrication of many material couples, but this is normally closed during subsequent rolling. If brittle intermetallics form at the laminate interfaces, delamination may occur during forging and rolling. However, with care, usable laminates can be rolled to sheet. Laminates which contain metals of very different mechanical properties undergo internal fracture during rolling, and this is shown to be the result of internal tensile stresses generated during the rolling operation.

Verbindungsschichtbildung beim Nitrieren und Nitrocarburieren

Only a limited understanding has been attained until now about the growth and constitution of the iron-(carbo)nitride compound layer at the surface of nitrided or nitrocarburized workpieces. Possible interpretations of well known phenomena are discussed. Equilibrium between nitriding/nitrocarburizing medium and the surface of the workpiece can be realized in practice only if certain conditions are obeyed. At relatively high temperatures and nitriding potentials stationary states can occur which are easily misinterpreted as states of equilibrium. The y' and E iron nitrides can be conceived as interstitial solid solutions exhibiting a more or less ordered arrangement of the interstitials, or might be considered as compounds, of a certain ideal stoichiometry, exhibiting a certain composition range. Diffusion in the iron nitrides was discussed in terms of diffusion of nitrogen over interstitial sites of the iron sublattice and of iron, too, if structural vacancies occur on the iron sublattice. Pore formation can always occur as consequence of the thermodynamic instability of the nitrides (N2 gas formation). Other, additional interpretations of pore formation are indicated: Kirkendall porosity and porosity induced by dissolution of iron (in salt baths). The enrichment of carbon in the bottom part of the compound layer is predominantly seen as a consequence of the gradient in carbon chemical potential in particular due to nitrogen-concentration depth profile.

NiAl formation by annealing of infiltrated aluminium-nickel precursors

Samples fabricated by pressure-infiltration of nickel powders with molten aluminium were heat treated at 1200 °C under vacuum or argon isostatic pressure. Reaction and diffusion in the asinfiltrated samples, which contained nickel, aluminium and varying amounts of Ni2Al3 and Al3Ni, resulted in the formation of NiAl as a principal phase with nickel and Ni3Al as minor phases. All samples exhibited macroporosity due to the formation of an interconnected transient liquid phase during heat treatment. Vacuum-annealed samples also showed extensive Kirkendall porosity in the nickel phase, which was, however, pore-free in hot isostatically pressed samples due to compaction during reaction. Concentration profiles of aluminium in these nickel regions were measured and are in good agreement with predicted values.

Properties of Reactively Synthesized Titanium Aluminides

AbstractReactive synthesis directly forms compounds from their elemental constituents. In this U.S. Bureau of Mines study the microstructure, hot-hardness and tensile properties of alloys and composites based on TiAl and Ti2AlNb were studied. Ti2AlNb sheets were processed from elemental Ti, Al and Nb foils. The elemental foils were consumed during processing, leaving a microstructure consisting of three phases: O (orthorhombic Ti2AlNb), B2 (ordered cubic Ti), and α2 (Ti3Al). The composite sheet structures were heat-treated to produce a microstructure consisting of O lathes in a B2 matrix. A significant advantage of reactive synthesis is the ease of formation of complex shapes prior to synthesis of the aluminide. TiAl-based alloys and in situ TiAl/boride and TiAl/Ti5Si3 composites were produced from elemental mixtures of Ti and Al with either additions of B or Si powders. The hot-hardness and tensile properties of these alloys and composites are reported. The Kirkendall porosity associated with reactive sintering of TiAl can be eliminated by simultaneously reactive sintering TiAl and Ti5Si3.

Degradation of TAB outer lead contacts due to the Au-concentration in eutectic tin/lead solder

The influence of the Au-concentration in OLB solder fillets on the contact reliability is shown. Quantitative analysis of the Au content shows a good accordance with the estimated concentrations using the geometrical data. The Au-concentration is the major influence factor on the thermal aging behavior of OLB contacts. A change in the failure mechanism due to Kirkendall porosity is observed if a Au-concentration of 9 wt% in the solder fillet is reached. The pores formed during diffusion at the interface between copper and the ternary intermetallic compound Cu/sub 3/Au/sub 3/Sn/sub 5/ cause a strong degradation of pull forces. For low Au-concentrations, the degradation of pull-forces and the growth of the ternary compounds CuAu/sub 5/Sn/sub 5/ and CuAu/sub 4/Sn/sub 5/ have comparable activation energies in the range of 0.3 eV. For higher Au-concentrations the two effects show different values. The activation energy of pull-force degradation increases significantly up to 0.67 eV, whereas the growth constant of the ternary compound shows a small increase to 0.4 eV. The investigations show evidence of a critical Au-concentration in OLB contacts. Kirkendall pore formation causes higher activation energies for the pull-test degradation compared to the activation energy for the growth rate of the ternary compound. >

An analysis of the critical conditions for diffusion-induced void formation in Ni-Cu laminate composites

An analysis that predicts, based on Kirkendall porosity formation, the critical composition difference that can exist without diffusion-induced porosity formation in laminate composites between solid-solution alloys is presented. Predictions based on laminate composites fabricated from pure nickel to nickel-copper solid-solution alloys indicate that porosity formation should be suppressed if the copper content in the solid-solution layer is below 40 at.%. The predictions are verified with metallographic observations on a series of annealed laminate composites. The implications of this study on the stability and mechanical properties of composites at high temperatures are discussed.

High temperature deformation and fracture behavior of laminate composites: The importance of interdiffusion during creep

Composites, as well as materials with solidified structures, exhibit composition and microstructures which systematically vary with position and mechanical properties which are often described with analyses based on the rule of mixtures. In addition to Kirkendall porosity formation, it is anticipated that interdiffusion during deformation at high temperatures will modify dislocation motion, or will modify dislocation structures. Therefore, the properties of as-cast weld metal or as-bonded laminate composites would differ from the same materials which received homogenization heat treatments to stabilize composition gradients, and simple analyses based on the rule of mixtures may not apply. In this paper the high temperature creep behavior of laminate composites with a constant volume fraction of each layer but with different layer thicknesses are evaluated under test conditions which produce dynamic composition gradients. The results of this study provide a physical basis for modeling the dynamic nature of composition gradients induced by processing (e.g. solidification, vapor deposition, roll bonding, etc.) and the effects of dynamic composition gradients on high temperature mechanical properties.

Formation of peripheral porosity regions around urania in zirconia-urania mixed oxide powder compact sintering

Sintering studies of zirconia-urania mixed oxide powder compacts (in stages of 5% urania up to a maximum of 20% addition) were carried out at temperatures between 1000–1400°C for various soaking periods. The formation of a peripheral porosity region around comparatively coarser urania particle was a characteristic feature in this mixed oxide sintered compact. At even a higher sintering temperature (1800°C), where extensive solid solution formation takes place, this porosity region demarcates the solutionized particles from the host zirconia apparently acting as a discontinuity in the system. Relative shrinkage difference between the dissimilar particles probably contributes to the porosity regions around the minor second phase at a lower temperature while at higher temperature generation of “Kirkendall porosity” may be responsible for such an effect.

Interdiffusion and coating design

Interdiffusion between a coating and its underlying substrate can change the oxidation properties of the coating and the mechanical properties of the coated material. An approach is given here for analyzing electron probe microanalyzer (EPMA) data in order to measure the amount of interdiffusion and to predict how interdiffusion can be reduced by changing the coating chemistry. The approach makes use of “zero flux planes” and “composition vectors”. It is shown how to eliminate or reduce Kirkendall porosity in a systematic manner. Experimental results obtained on nickel base alloys are used for illustration.

An analysis of diffusion-induced porosity in CuNi laminate composites

Nucleation, growth, migration and coalescence of diffusion-induced pores in CuNi laminate composites were observed. The pore structures were interpreted based on an analysis of Kirkendall porosity formation

The effect of composition on marker movement and kirkendall porosity in ternary alloys

A linear model which predicts Kirkendall marker movement in multicomponent diffusion couples has been tested with a series of Ni-Cr-Al γ-phase alloys. The test results followed the model to within experimental scatter, even though the model assumes constant diffusivity in the reaction zone. Comparing the model predictions with previous work suggests that the model can be used in alloy design to eliminate Kirkendall porosity. In addition, preliminary results suggest that the coefficients of the model can be predicted with useful accuracy from fundamental diffusion data.

A model for the Kirkendall effect

A model is proposed which predicts that Kirkendall marker movement is a linear function of alloy composition. The model assumes that concentration gradients and intrinsic diffusivities at the Kirkendall plane are similar to those at the Matano plane; and that concentration gradients can be approximated by referring to a constant diffusivity system. The physical constants in the model are functions of intrinsic diffusivities and square root diffusivities. The model suggests how Kirkendall porosity can be reduced or eliminated in ternary and higher order alloys by proper alloy selection.

Diffusional transport during the cyclic oxidation of ?+?, Ni-Cr-Al(Y, Zr) alloys

The cyclic oxidation behavior of several cast γ+β, Ni-Cr-Al(Y, Zr) alloys and one LPPS γ+β, Ni-Co-Cr-Al(Y) alloy was examined (γ, fcc; β, NiAl structure). Cyclic oxidation was performed by cycling between 1200°C and approximately 70°C. Oxide morphologies and microstructural changes during cyclic oxidation were noted. Recession of the high-Al β phase was nonparabolic with time. Kirkendall porosity resulting from diffusional transport within the alloy was observed in the near-surface γ-phase layer of one alloy. Concentration profiles for Ni, Cr, and Al were measured in the γ-phase layer after various cyclic oxidation exposures. It was observed that cyclic oxidation results in a decreasing Al concentration at the oxide-metal interface due to a high demand for Al (a high rate of Al consumption) associated with oxide scale cracking and spalling. In addition, diffusion paths plotted on the ternary phase diagram shifted to higher Ni concentrations with increasing cyclic oxidation exposures. The alloy with the highest rate of Al consumption, and highest Al content, underwent breakaway oxidation after 500 1-hr cycles at 1200°C. Breakaway oxidation occurred when the Al concentration at the oxide-metal interface approached zero. The relationship between the Al transport in the alloy and breakaway oxidation is discussed.